基于 Copilot SDK 构建的 GitHub Copilot 原生桌面图形界面。

先修条件

• Node.js 22+
• GitHub Copilot订阅
• GitHub CLI 已安装并认证 — 如果您还没运行，请运行 gh auth login

安装

那么，克隆仓库：

这会安装依赖、运行测试、构建应用并打包。输出会发送到 release/ 目录：

Windows： 在 npm run dist 之前运行 pwsh -NoProfile -File .\scripts\setup-windows.ps1 以设置构建依赖。

Linux： 在 npm run dist 之前运行 sudo ./scripts/install-linux-deps.sh 以安装原生依赖。

特色

• 🗂️ 标签会话 — 多个对话，每个对话都有自己的工作目录和模型
• 🌳 Git 工作树会话 — 将 GitHub issue URL 粘贴→独立的工作树 + 会话
• 🔁 Ralph Wiggum — 迭代代理模式：设置完成条件，循环直到完成

Cooper

A native desktop GUI for GitHub Copilot, built on the Copilot SDK.

Prerequisites

• Node.js 22+
• A GitHub Copilot subscription
• GitHub CLI installed and authenticated — run gh auth login if you haven’t already

Installation

Clone the repo, then:

npm run dist

This installs dependencies, runs tests, builds the app, and packages it. The output goes to the release/ directory:

Windows: Run pwsh -NoProfile -File .\scripts\setup-windows.ps1 before npm run dist to set up build dependencies.

Linux: Run sudo ./scripts/install-linux-deps.sh before npm run dist to install native dependencies.

Features

• 🗂️ Tabbed Sessions — Multiple conversations, each with its own working directory and model
• 🌳 Git Worktree Sessions — Paste a GitHub issue URL → isolated worktree + session
• 🔁 Ralph Wiggum — Iterative agent mode: set completion criteria, let it loop until done
• 💻 Embedded Terminal — Real PTY per session, one click to attach output to your prompt
• 🎤 Voice Input/Output — Speech-to-text and text-to-speech
• 🔌 MCP Servers — Model Context Protocol for extended tool capabilities
• 🎯 Agent Skills & Subagents — Personal and project skills via SKILL.md, .agent.md, and subagent delegation
• 🤖 Model Selection — Dynamic model list with favorites — GPT-5.2, Opus-4.6, Sonnet, Haiku, Gemini, and more
• ⚡ YOLO Mode — Auto-approve tool calls for uninterrupted agent flow

Development

npm install && npm run dev

See CONTRIBUTING.md for details.

Community

Join our Discord to report bugs, request features, and chat.

License

MIT